There has been a rapid expansion in treatment options for the management of metastatic prostate cancer, but individual patient outcomes can be variable due to inter-patient tumor heterogeneity. Fortunately, the disease can be stratified on the basis of common somatic features, providing potential for the development of clinically useful prognostic and predictive biomarkers. Tissue biopsy programs and studies leveraging circulating tumor DNA (ctDNA) have revealed specific genomic alterations that are associated with aggressive disease biology. In this review, we discuss the potential for genomic subtyping to improve prognostication and to help guide treatment selection. We summarize data on associations between AR pathway alterations and patient response to AR signaling inhibitors and other standards of care. We describe the links between detection of different types of DNA damage repair defects and clinical outcomes with targeted therapies such as poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitors or immune checkpoint inhibitors. PI3K signaling pathway inhibitors are also in advanced clinical development and we report upon the potential for these and other novel targeted therapies to have impact in specific molecular subsets of metastatic prostate cancer. Finally, we discuss the growing use of blood-based analytes for prognostic and predictive biomarker development, and summarize ongoing prospective biomarker-driven clinical trials.